Created 12 November 2025
McClendon, Sperber, Kaaba BJPS
Created and run on Stata version SE 2018, on macOS Sequoia 15.7.1

*****Instructions*********
To reproduce most results in the study, run "SperberKaabaMcClendon_genderpaper.do" in Stata. Analyses described in the main paper and the appendix are labeled by Table and Figure number in the .do file. The .do file calls all .dta files described below. The .do file was created on Stata version SE 2018 and does not require additional packages or libraries.

See "Replication Data for: Comparing Religious and Secular Interventions to Increase Young Adult Political Participation: Evidence from WhatsApp-Based Civic Education Courses in Zambia" at https://doi.org/10.7910/DVN/UBU601 for files to create the cleaned homework answers datasets below from the raw Twilio data.


*****Summary of files to reproduce main results in paper and appendix*****
1. SperberKaabaMcClendon_genderpaper.do:
Stata Do-file containing code for merging relevant files, for coding variables, and for replicating analyses reported in the main paper. Analyses for the main paper are labeled by Table and Figure number, as they appear in the main paper.

2. pp_baseline_cleaned.dta, pp_endline_cleaned.dta and pp_followup_cleaned.dta:	
Stata data files supplied by IPA Zambia to the authors with responses to baseline, endline, and follow-up phone surveys. IRB/UNZA Research Ethics required that the authors receive only de-identified data.
	
3. ccmg_behavioral.dta: Stata data file supplied by CCMG-research team liaison, which lists all study participants (by caseID; respondent email addresses have been removed) who signed up for the listserv and tracks whether they responded to email invitations and signed up for pvt and data analysis activities.\

4. post_election_survey.dta: Stata file containing responses to WhatsApp survey done in the days after the election.

5. homeworks_count.dta: Stata file containing word counts of open-ended responses to homework questions, derived from clean_Twilio.dta.

6. homeworkanswers_identity.dta: Stata file containing handcoded indicators for whether respondents indicated religious, youth, gender, ethnic identities in the open-ended responses to homework questions; coded by PhD RA.

7. hashtagposts.dta: Stata file containing indicators for whether respondents who provided FB or Twitter accounts used the study-provided hashtag (not many at all). De-identified.

8. poltweets_collapsed.dta: State file with number of tweets of different kinds per respondent id (caseID). De-identified.

9. FB_compiled.dta: Stata file with indicators for whether each supplied Facebook and instagram handle used the hashtags YouthVote2021, ZambiaDecides, CanYouRemember, mentions voting, or Hakainde Hichilema (HH). De-identified. Organized by caseid (respondent number).

10. twitter_connections.dta: Stata file with numbers of twitter connections with other participants for participants with twitter handles who had other connections, with share of those connections across assigned tracks. De-identified. Only 19 relevant study participants.

Additional files for separately produced figures:
- "Figs 1 and 2 gender descriptive stats with pol interest.xlsx": an Excel file with the values and figures of bar graphs in Figures 1 and 2 in the main paper (see main .do file for code used to generate values)
- "Fig6_women_individualactions.R": R code to generate Figure 6 in the main paper (see main .do file for code used to generate values)
- "Figure 7 volunteer gender gap.R" : R code to generate Figure 7 in the main paper (see main .do file for code used to generate values)
- "Zambia_R9_2022": Stata code (version 18) to generate Figure C.1 in the appendix using publicly available Round 9 data from the Afrobarometer.
